ComfyUI Power Tips - Save time with stable diffusion so you can sleep!

I've come across interesting new ways to speed up ComfyUI workflow and content creation that will hopefully help you save some time!

RESOURCES

KEY STABLE DIFFUSION COMFYUI TOPICS:
* Node Templates
* Impact Bridget
* Updated Cut / Paste / Blend
* Image 2 Image
* Clipseg Masking

[beveresmoor]

Comfy UI is so versatile that I am still learning what I can do with it. Thank you for showing this cool tip. 😊

[mtdong]

Excellent video! So concise with a great mix of ui navigation and workflow optimization tips. 🙌

[cyril]

thanks for the impact bridge, I was looking for that for a while!

[MajorStudioArt]

Impact Bridge is super cool; I might start using that as a simpler alterative to the ways I'm currently handling switches, especially the bypyassers.

[jgmemxj]

you video is amazing, these nodes are definitely what I wanted, It's super helpful to create my workflow, thank you❤

[blender_wiki]

Nice video for beginners however is interesting how YouTubers are always worried about speed and don't care about quality. In a creative workflow you, the human, you are always the slowest element and having a super speed tool doesn't make you art better but gives you only more anxiety.
Furhermore Instead of bridge is often more useful using a switch(or more) that you control with a set/get 'remotely'.

[MajorStudioArt]

Some important caveats about templates: They are copy/pasted, and as such there are some things that I've found that won't work. One, groups can't be stored in a node template. Two some custom nodes might not work when brought in as a node template---specifically, currently RGThree's amazingly useful node bypassers and bypass repeaters have limited functionality past two inputs (when used within a template)

[chillsoft]

Very cool! How did you get that "fullscreen" sidebar? Looks much cleaner! Thanks for the video btw, the blocker thing is essential!

[Martin-bxet]

When I open the workflow there is an error message that the 'IPAdapterCLIPVisionEncode' is missing and the node at the 'Clip Vision Encode' position shows as undefined in red.

I can add a (standard ?) 'Clip Vision Encode' node but when I swap it in for the red node, the 'Apply IPAdaptor' node errors and reports that required inputs 'clip_vision' and 'image' are missing. Does this mean I am adding an invalid 'Clip Vision Node' or is something else amiss?

(The blonde woman with the coffee cup shows up okay in the preview bridge and I am selecting the appropriate pytorch_model into the 'Load CLIP vision' node.)

[hindihits]

how to get the milehigh styler? can't seem to find it...

[pfbeast]

I just try to your mask using words workflow, but i use this wokflow,
1. In apply ipadapter node 'clip_vision_output' input not present, in your video in apply ipadapter have total 4 input (model, ipadapter, clip_vision_output & mask), but in my apply ipadpater node have total 5 input (ipadpater, clip_vision, image, model, attn_mask) present, not present 'clip_vision_output' input. I have uptodate ip-adapter plus node & model, still not present.

2. in my comfyui models folder, 'clipseg' folder not present and CLIPSeg Model Loader not load any model. now tell me what should i have to do?